strtok ? ?

[Peter]

hi all, the strtok() cannot phrase the token within another token, am i
correct?

For example, i want to get the second word of every row of a file, how
to use strok to complete this?
thanks
from Peter (cm****@hotmail .com)

[Tilman Kuepper]

Hi Peter,

For example, i want to get the second word of every row of a file, how
to use strok to complete this?

You might also consider a stream-based solution:

#include <string>
#include <sstream>
#include <fstream>
#include <iostream>

int main()
{
std::ifstream f("ReadMe.txt") ;

// Read all lines from test.txt...
std::string line;
while(std::getl ine(f, line))
{
// Read 2nd word...
std::istringstr eam tmp(line);
std::string first, second;
tmp >> first >> second;

std::cout << second << "\n";
}
}

Best regards,
Tilman

[Puppet_Sock]

Peter wrote:

hi all, the strtok() cannot phrase the token within another token, am i
correct?
I presume you mean "parse."

If by "another token" you mean the delimeter, then yes, you need the
string to be parsed, and the delimeter to parse it.
For example, i want to get the second word of every row of a file, how
to use strok to complete this?

You decide what delimeter to use to divide the string into words.
Presumably a space. Then you read the file line by line, and
feed each line to strtok, with the space as delimeter, and call
it appropriately to get the second item. With appropriate code
around the call to take care of such things as there not being
two words on a line. You also decide what to do about such
weird cases as "my car,house,boat" etc., where there is some
strange thing with missing space. And so on.

Also, strtok has one really good use. It teaches newbies what
not to do, by biting them in the ass. Get hold of a good description
of strtok and read up on such things as: Buffer over runs as a
substantial security risk. Static buffers shared between different
calls to strtok as a data corruption risk. And so on.

When all of that gets on your nerves adequately, you think about
a stream based solution.
Socks

[Default User]

Peter wrote:

hi all, the strtok() cannot phrase the token within another token, am
i correct?

For example, i want to get the second word of every row of a file, how
to use strok to complete this?

Automatically? No. It's a fairly easy algorithm to write.

Generally, strtok() is not recommended. It works only on C-style
strings. It modifies the string. It can't be used with constant strings
or string literals. It has concurrency and reentrancey problems.

If all you need to do is get the second word of string, using
std::string makes a lot more sense.

Here's a little something I ginned up. Tested for functionality. May
have problems, review carefully before use.

#pragma warning( disable : 4786)

#include <vector>
#include <string>
// retrieves the nth substring, returns:
// 0 on success,
// -1 for invalid number
// 1 for insufficient number of substrings

int GetNthString (const std::string &inString,
const std::string delimiters,
std::string &outString,
int n)

{
int retVal = 0;
std::string tmpString;
std::string::si ze_type start = 0;
std::string::si ze_type end = 0;
bool done = false;
int i = 1;

if (n <= 0)
{
return -1;
}

while (!done)
{
end = inString.find_f irst_of (delimiters, start);
if (end == std::string::np os) // end of string
{
done = true;
}

if (i == n)
{
outString = inString.substr (start, end-start);
done = true;
}

else
{
if (done)
{
retVal = 1;
}
else
{
start = inString.find_f irst_not_of (delimiters, end);
i++;
}
}
}

return retVal;
}
#include <iostream>
using namespace std;

int main()
{
string s1 = "one and two and three";
string s2;
int res;

res = GetNthString (s1, " \t", s2, 2);

if (res == 0)
{
cout << s2 << endl;
}

else
{
cout << "Error: " << res << endl;
}

return 0;
}

[Default User]

Default User wrote:

Here's a little something I ginned up. Tested for functionality. May
have problems, review carefully before use.

#pragma warning( disable : 4786)
Urg. Didn't mean to paste that in. Can you tell I use VC++ 6? Sorry gnu
people.
#include <vector>

This wasn't needed, left over from the program I savaged to make the
new one.

Brian

[Peter]

Sorry, forgive my bad english.
What i mean is,if i want to get the second word of every delimiter ":",
i cannot use strtok, right?

For example, here is my testing char *:
abc,def,hij:pet er,john,susan:1 23,456,789

If i use strtok to separate the string into:
abc,def,hij
peter,john,susa n
123,456,789
Then i cannot use strtok to parse again on each substring.

And i want to get every second word of every ":", how to? suppose the
answer is def, john and 456.
In my os library, i only have strtok, and i found strtok is not working
in this case.
thanks
from Peter (cm****@hotmail .com)

[Peter]

Thank you all guys.
I really need a string class in my os.
thanks again

[Default User]

Peter wrote:

Sorry, forgive my bad english.
What i mean is,if i want to get the second word of every delimiter
":", i cannot use strtok, right?

For example, here is my testing char *:
abc,def,hij:pet er,john,susan:1 23,456,789

If i use strtok to separate the string into:
abc,def,hij
peter,john,susa n
123,456,789
Then i cannot use strtok to parse again on each substring.
Why not? Just start a new strtok sesssion (so to speak) with the
pointer to the substring you want to use and the new delimiter.
And i want to get every second word of every ":", how to? suppose the
answer is def, john and 456.
In my os library, i only have strtok, and i found strtok is not
working in this case.

It's not normally the "os libary" that is of interest, it's the
standard library of the implementation. An OS library may be used by
the implementation, of course.

What compiler are you using? std::string was in the 1998 C++ standard,
and most compilers of recent vintage provide it.

You will be better off if you use it with something like the code I and
another person posted yesterday, rather than using the rather tricky
strtok().

Brian